New materials for DUAL:

the LNL activity

A

B

B

Α

B

Α

2

Y

Y

S

S

hh

hh

Dual detector : Best material parameters

• Two different materials ‘A’ and ‘B’ with two different Young modulus Y and density

•Sensitivity curve optimized in the same frequency window

• QL readout

•Toroidal shape

•No thermal noise

BΑ,hhS

BΑ,

BΑ,YStrain PSD of material A,B Young modulus of mat. A,B

Density of material A,B

/2Y

Dual detector : Candidate Materials

810/ TQ

-Diamond 163 ? expensive

-Silicon carbide 27 Not well known available in large size

-Berillium 24 <106 expensive

-Sapphire 13 108 expensive+need bonding

-Molibdenum 6.8 107 available in large size

-Silicon 4.4 >108 need bonding

-others ? (i.e. Alumina (17), MoCu (), CuBe)

5056/ Alhh

Mathh SS Q

1. Minimize sensitivity curve i.e. maximize

2. Minimize thermal noise

Red= Material under investigation

Sintered silicon carbide: results•Cantilevers of different thickness (0.3-0.5 mm) and lenght (5-10 cm)

•Both optical lever and capacitive readout

Similar results in E.K. Hu et al. Phys Lett. A 157, 209 (1991) -> Annealing should improve the Q

0 50 100 150 200 250 3004E-5

1E-4

8E-4

Loss

ang

le

Temperature [K]0 50 100 150 200 250 300

5E-5

1E-4

1E-3

Loss

Ang

le

Temperature [K]

Sintered silicon carbide: results

0 50 100 150 200 250 3005E-5

1E-4

1E-3

Loss

Ang

le

Temperature [K]

Sintered SiC Annealed 1900 C Annealed 1200 C

0 50 100 150 200 250 3004E-5

1E-4

8E-4

Loss

ang

le

Temperature [K]

Sintered SiC Annealed 1900 C Annealed 1200 C

•Cantilevers of different thickness (0.3-0.5 mm) and lenght (5-10 cm)

•Both optical lever and capacitive readout

Similar results in E.K. Hu et al. Phys Lett. A 157, 209 (1991) -> Annealing should improve the Q -->Seem not very much

Infiltrated silicon carbide C-SiC: results

•Best achieved loss angle 2x10-6

•Annealing did’nt improve the quality factor

•Measured samples: two cantilever of different thickness and lenght from Cesic (Germany)

•Different Carbon matrices

•Capacitive readout

Annealed

Not Annealed

0 50 100 150 200 250 3001E-6

1E-5

1E-4

1E-3

Loss

ang

le

Temperature [K]

I nfiltrated SiC thickness 5 mm 1030 Hz Infiltrated SiC thickness 3 mm 977 Hz

0 50 100 150 200 250 3001E-5

1E-4

1E-3

Loss

Ang

le

Temperature [K]

I nfiltrated SiC thickness 3mm Same annealed at 1900 C

Silicon samples: bonding research

•Ready made for many bond proccesses (Anodic bonding, Eutetic bonding,Adhesive bonding,Fusion bonding,Thermocompression bonding)

•Wafer diameter 100 mm

•Stack thickness 6 mm

Machine capabilities

A Silicon Wafer bonder is now availabe at the

Mt-Lab in Trento

Silicon samples: bonding loss angle

Bonding LayerSilicon wafer

Bonded silicon disk h=0.9 mm d=100 nm

)()( bondtot

bondbondednot E

E

Disk loss angle

Ansys Analysis

SS pistonSS spring

Sapphire balls 1mm

Si Disk

Not in scale cross section

Si disk suspension set-up

Al

Silicon <100>

<111> crystal plane side wall

TMAH

175 m deep pyramid hole, with square opening 500 m side

silicon bulk wet etchingSet up #1

SS springSS piston

Sapphire balls 1mm

Si Disk

Not in scale cross section

Si disk suspension set-up

Al

Hole about 300 m in diameter , Milled using dentist’s tools

Set up #2

Si disk displacement Readout

Laser

Quadrant photodiode

Achieved sensitivity

10-8 -10-9 m

Observed warm-up effects at low temperature !

Thus not ideal for ultracryogenic operations

Si disk displacement Readout

Rotate View

Capacitive readout

Vbias

Vout

Si disk: capacitive redout

Readout sensitivity

Achieved Sensitivity during run #1

(Vbias=60 Volt, gap= 0.1 mm)

][10][ 2 nmxVoltVout

High sensitivity require

00 /1 CC

x

dx

dC

C

VV

Par

biasout

Vbias

C(x)

•Small gap

•Low parasitic capacitance (Cpar)

•Low noise voltage preamplifiers (SQUID amplifer can improve sensitivity very much)

The comb capacitor capacitance value C(x) is a function of the distance (gap+x) between them and the opposite dielectric plate

Si disk: first cryogenic run set-up

• Si <100> oriented Boron doped

•Disk diameter 4 inch

•Thickness 0.5 mm

Misalignement problem

Adopted suspension (not optimized) set-up

Si disk

Sapphire balls

SS piston

2mm

Si disk: first cryogenic run results

0 50 100 150 200 250 3001E-7

1E-6

1E-5

1E-4

Mode freq 980 HzMode freq 997 Hz

Loss

Ang

le

Temperature [K]0 50 100 150 200 250 300

1E-7

1E-6

1E-5

1E-4

Mode freq 383 Hz

Loss

Ang

le

Temperature [K]0 50 100 150 200 250 300

1E-7

1E-6

1E-5

1E-4 Mode freq 484 Hz

Loss

Ang

le

Temperature [K]

First Si Disk cryogenic run: Quality factor limitations

• The contribution of thermoelastic damping and surface losses at 4.2 K should be less then 10-8

•We are presumably dominated by suspension losses and/or sample microfracture induced by manufacturing the central hole

•However for this specific run gas damping should play a relelvant role because we had a cold leak

mBarPmm

Kg/m4

Mode3

Si

1015.0H

Hz483ν2300ρ

P

1

M

RTνHρ

2Q

HeModeSi

3/2

Christian’s Model

8105.2Q

L

16

P

1

M

RTνHρ

2Q

HeModeSi

3/2Gap

mm

mm

35L

1.0Gap

7106.3Q Comb

Capacitor

Bao et al. Model (sqeezed film damping )

Even worse using true gas dynamic models and outgasing

Conclusions

•SiC is very interesting for its high sound velocity, but the sintered and the infiltrated silicon carbide, that can be fabricated in large size as required for the Dual detector, show at low temperature a quality factor that is at least 2 order of magnitude lower than the required value. The quality factor of monocristalline SiC should be better but in this case we have to develop low losses bonding procedures and take care of the cost.

•Monocristalline Silicon is also a candidate material but it is not available at the required size. We plan to measure the Si bond losses to evaluate if a big elastic body can be obtained bonding togheter smaller pieces without affecting the overall Q fator

•Molybdenum is the best candidate material for the Dual detector however we are considering other material considering other materials as MoCu, pure Alumina, CuAl